[OSM-talk] Highway=trunk : harmonization between countries ?
dieterdreist at gmail.com
Tue Aug 22 19:35:30 UTC 2017
sent from a phone
> On 22. Aug 2017, at 13:29, Colin Smale <colin.smale at xs4all.nl> wrote:
> I don't understand why junction penalties should be dependent on the road classes, and not on physical characteristics. I guess this is just a heuristic which can be useful if you don't have the full picture.
often on junctions passing on the higher road will be advantageous compared to approaching via lower roads (which typically have to give way, or will have shorter green light, etc.). That's the classic situation which is probably the reason that higher rated roads are preferred by routing engines.
It is of course a generalization, and in other (probably much fewer) cases, the main road might have always green light but the other roads will get green light when a car is approaching (induction loops, etc.). When the junction is a roundabout the road class won't matter. It's really complex to get "the full picture" of all real world constellations that do exist, even more with technology interacting with traffic lights, ideally you would have to have real time information about other vehicles and about traffic light states in order to choose the current best solution (even better would be predicting the traffic for the time when you will be there, surely quite complex task).
I don't say prioritizing bigger highway classes is the absolutely best thing to do, but it's probably for most cases a good solution and far less complicated to calculate.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the talk